Audio clip and spectrogram of a real song:¶
In [8]:
ipd.Audio(shire[0:ytime], rate = sr)
Out[8]:
In [9]:
a = len(t)/5
plt.figure(figsize=(18,10)); plt.imshow(np.abs(S[0:maxN,:]), cmap='inferno',vmin=0, vmax=0.01, origin='lower', extent=(0,S.shape[1],0,freqmax))
plt.xticks((0,a,2*a,3*a,4*a,5*a), ('0:00','0:30','1:00','1:30','2:00','2:30')); plt.xlabel('Time [minutes]'); plt.ylabel('Frequency [Hz]'); plt.title('Spectrogram of Shire'); plt.show()
But why not start from a picture to create audio?¶
In [10]:
imageToAudio('diagonal2.png')
Out[10]:
In [11]:
imageToAudio('derpcorn.png')
Out[11]:
In [12]:
imageToAudio('banksky.jpg')
# https://www.sothebys.com/en/articles/sothebys-gets-banksyed-at-contemporary-art-auction-in-london
Out[12]:
In [13]:
imageToAudio('trump.jpg')
Out[13]:
In [14]:
imageToAudio('monalisa.jpg')
Out[14]:
